A virtual hotel tour customization system based on intelligent optimization and multimodal interaction

Abstract

The present invention discloses a virtual hotel tour customization system based on intelligent optimization and multimodal interaction, relating to the field of virtual reality and augmented reality application technologies. The system includes a multi-source data acquisition module, a VR tour module, an AR customization module, a user interaction interface, and a data storage management module. The multi-source data acquisition module acquires hotel 3D data and user behavior data, the VR tour module plans user movement paths, the AR customization module adjusts room layouts based on user behavior and selects the optimal solution, and the user interaction interface provides touch and voice interaction. Through intelligent optimization and multimodal interaction technology, the present invention achieves personalized tours and room layout customization, improves user experience, plans the optimal path, intelligently adjusts the layout based on user behavior, provides a user-friendly touch and voice interaction interface, and uses encrypted storage technology to ensure data security, bringing innovative solutions to the field of virtual reality and augmented reality application technologies.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual reality and augmented reality applications, and in particular to a virtual hotel tour customization system based on intelligent optimization and multimodal interaction. Background Art

[0002] A virtual hotel tour is a travel guide service that combines virtual reality and augmented reality technologies. It uses head-mounted displays, sensors and other devices to bring users into a computer-generated three-dimensional hotel environment for an immersive tour experience. Through the virtual hotel tour, guests can intuitively understand the hotel's internal structure, room layout, facility location and surrounding attractions before checking in. This tour method not only provides a more vivid and specific display effect, but also allows guests to feel as if they are there, experience the hotel atmosphere in advance, choose their favorite room or explore the activity area they are about to participate in. In addition, the virtual hotel tour can also integrate online booking functions, making it convenient for guests to directly arrange their activity schedules, greatly simplifying the travel planning process. With the popularization and integration of related technologies, virtual hotel tours will provide customers with more personalized and interactive tourism information services, becoming an indispensable part of smart hotels.

[0003] To address the weak immersion and poor interactivity of traditional virtual hotel tours, existing technologies use basic 3D modeling and simple interactive interfaces. Although this approach can provide a certain degree of visual display, allowing users to browse every corner of the hotel in a virtual environment, it can also result in abrupt scene switching and limited interactive elements. During the tour, users are often restricted to browsing along preset routes and cannot freely explore or deeply interact with the environment. This lack of real immersion and participation leads to a poor user experience and prevents them from fully experiencing the hotel's unique charm and service details. This also affects the hotel's ability to attract potential customers and enhance its brand image through virtual tours. Therefore, a virtual hotel tour customization system based on intelligent optimization and multimodal interaction is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a virtual hotel tour customization system based on intelligent optimization and multimodal interaction to solve the problems raised in the above background technology.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is: a virtual hotel tour customization system based on intelligent optimization and multimodal interaction, including a multi-source data acquisition module, a VR tour module, an AR customization module, a user interaction interface and a data storage management module;

[0006] The multi-source data acquisition module collects the hotel's 3D spatial point cloud data and user customized behavior data, and performs preprocessing and feature extraction on them;

[0007] The VR navigation module uses a bounding box-based collision detection algorithm based on 3D spatial point cloud data to calculate the positional relationship between the user's virtual character position and each bounding box, and plans the optimal movement path;

[0008] The AR customization module selects and adjusts room layout elements based on user customization behavior data, calculates the layout score of each plan, and selects the best layout plan;

[0009] The user interaction interface generates a touch and voice interaction interface that displays VR tour results and AR customization effects;

[0010] The data storage management module encrypts the hotel VR tour data and AR customization data, and uses big data storage and management technology to store the encrypted data.

[0011] A further improvement of the technical solution of the present invention is that the multi-source data acquisition module, the hotel's 3D space point cloud data and user customized behavior data acquisition process includes:

[0012] LiDAR sensors are deployed in the hotel lobby, corridors, and at the center of the guest room ceilings, covering a 360-degree horizontal field of view and a 60-degree vertical field of view at a fixed downward angle. Depth cameras are deployed at the four corners of guest rooms and on pillars in public areas, 2.5 meters above the ground, with a vertical field of view covering the entire height of the room and a 90-degree horizontal field of view. LiDAR emits a pulsed laser beam, measures the time difference between emission and reflection, calculates the distance to the surface of objects, and generates 3D point cloud data containing three-dimensional coordinates and reflection intensity information for each point. Depth cameras use structured light projection to capture the pattern offset after surface deformation and calculate depth values ​​based on triangulation principles. This generates dense 3D point cloud data with a resolution of 0.5 cm to supplement furniture edge details.

[0013] The touch terminal is embedded in the hotel's official website and mobile application interface to support touch-screen operation. It collects the RGB values ​​of the user's selected bedding color and furniture translation instructions. The voice collection module is integrated into the mobile device microphone and VR headset to receive voice instructions in real time and distinguish the user's identity through voiceprint recognition. When the user touches the screen to select a color block, the virtual hotel tour customization system extracts the RGB value corresponding to the screen pixel coordinates to collect the user's color selection data. When the user drags the virtual furniture model, the virtual hotel tour customization system records the coordinate difference between the starting point and the end point to obtain layout adjustment instructions.

[0014] A further improvement of the technical solution of the present invention is that: in the multi-source data acquisition module, the process of pre-processing the collected 3D space point cloud data and user customized behavior data includes:

[0015] Based on the principle of statistical filtering, the average distance and standard deviation between each point in the point cloud and its neighboring points are calculated. If the average distance between any point and its neighboring points exceeds the average distance plus three times the standard deviation, it is determined to be an outlier and removed. The adjacent complete wall point cloud data is used to complete the missing wall area through plane equation fitting. In the pre-processed 3D point cloud, the path nodes of the passable area are marked and a cost value is assigned to each node. The denoised point cloud data is merged with the completed wall data to generate a 3D spatial model of the virtual hotel. This is input into the collision detection and path planning algorithm of the VR guide module. The cost value table of the path nodes is transmitted to the path planning algorithm as the weight basis for calculating the movement path.

[0016] The RGB value selected by the user is converted into a value in the range [0, 1] through normalization. The coordinate difference between the starting point and the end point is converted into absolute coordinates based on the room coordinate system. The normalized color data and absolute coordinate layout instructions are associated with the user account, encrypted and stored, and input into the AR customization module.

[0017] A further improvement of the technical solution of the present invention is that: in the multi-source data acquisition module, the process of extracting features from the pre-processed 3D space point cloud data and user-customized behavior data includes:

[0018] The preprocessed 3D point cloud data was projected onto a horizontal plane and divided into grids with a side length of 0.1 meters. The number of grids covering the virtual hotel wall point cloud was counted and multiplied by the area of ​​each grid to obtain the total area of ​​the room. The plane distance between the center points of each piece of furniture was calculated based on the vertex coordinates of the furniture bounding box. The length, width, and height were calculated based on the vertex coordinates of the obstacle bounding box. In the point cloud of the traversable area, the ground center point was extracted as the path node, with a spacing of 0.5 meters between adjacent nodes. Based on the attributes of the area where the path node is located, a cost value of 1 for corridor nodes and a cost value of 2 for stair nodes were assigned. These values ​​were used as weight parameters in the path planning algorithm.

[0019] Count the translation vectors in the user's historical adjustment data, calculate the high-frequency adjustment direction and amplitude, classify the user adjustment data using the K-means clustering algorithm, and extract the translation vector corresponding to the cluster center;

[0020] The total area of ​​the room, the distance between the center points of furniture, and the size of obstacles are input into the collision detection and path planning algorithm of the VR navigation module, and the layout adjustment features are input into the AR customization module.

[0021] A further improvement of the technical solution of the present invention is that: in the VR navigation module, the process of using a collision detection algorithm based on a bounding box to calculate the position relationship between the user's virtual character position and each bounding box includes:

[0022] Extract the vertex coordinates of the wall and furniture from the pre-processed 3D space point cloud data, calculate the minimum and maximum values ​​of their three-dimensional coordinates, and form an axial bounding box;

[0023] The user moves through the VR device. The virtual hotel tour customization system obtains the avatar's three-dimensional coordinates in real time based on the VR device's positioning system. It traverses the bounding box in the scene and determines whether the avatar's current position overlaps with the bounding box's spatial range. If the avatar's three-dimensional coordinates of the x-axis, y-axis, and z-axis are all within the minimum and maximum values ​​of the bounding box's three-dimensional coordinates, the avatar is determined to have collided with the bounding box.

[0024] If a collision is detected, the virtual hotel tour customization system will enforce constraints based on the bounding box type, prohibiting the virtual character from entering the wall bounding box area, correcting the virtual character's coordinates to the nearest edge outside the bounding box, allowing the virtual character to approach the furniture bounding box area but restricting penetration, and triggering a red light flashing prompt. The bounding box type includes walls and furniture.

[0025] A further improvement of the technical solution of the present invention is that: in the VR navigation module, the process of planning the optimal movement path from the user's starting point to the search end point includes:

[0026] Receive path node coordinates, proxy value table, user start and end point coordinate data through the VR navigation module;

[0027] The actual cost is obtained by adding the cumulative cost value from the starting point to the current node according to the path node cost value. The straight-line distance from the current node to the end point is calculated using the Euclidean distance to obtain the estimated cost. The A* algorithm is used to define the total cost function as the sum of the actual cost and the estimated cost.

[0028] The starting point is added to the open list, its total cost function value is recorded, and an iterative search is performed. The iterative search process includes selecting the node with the smallest total cost function value from the open list as the current node. If the current node is the end point, the path is backtracked and output. If the current node is not the end point, its adjacent nodes are traversed. For each adjacent node, the actual cost and estimated cost are calculated, and the total cost function value is updated. If the adjacent node is not in the open list and there is a better path, the current node information is updated and added to the open list. The iteration is terminated until there is no feasible path, and the optimal path node sequence is output. The optimal path node sequence is transmitted to the VR scene rendering engine, displayed in real time with green lines, and the node cost value is dynamically adjusted according to the hotel traffic.

[0029] A further improvement of the technical solution of the present invention is that in the AR customization module, the process of matching the user-selected bedding color using a color recognition algorithm based on the RGB color space includes:

[0030] The normalized RGB value of the user's selection is matched with the bedding color library of the virtual hotel. The Euclidean distance between the user's color and the colors in the bedding color library is calculated. The color with the smallest distance is selected as the matching result, and its RGB value is recorded as the user preference feature. The RGB value of the best matching color is associated with the user account, encrypted and stored in the data storage management module, and the matched bedding color is superimposed on the virtual guest room model in the AR scene in real time.

[0031] A further improvement of the technical solution of the present invention is that: in the AR customization module, a simulated annealing algorithm is used to calculate the scores of each furniture layout plan, and the process of selecting the best furniture layout plan includes:

[0032] Generate an initial furniture layout plan based on the pre-processed 3D room model, furniture size data, and the user's historical layout preference characteristics. Define furniture positions using 3D coordinates and orientation angles, comprehensively analyze the distance between furniture and the wall, and calculate the furniture layout plan score.

[0033] Apply position perturbations to random furniture in the current layout and calculate the score of the new furniture layout solution If the score of the new furniture layout plan is lower than the score of the current furniture layout plan , then accept the new furniture layout plan, otherwise accept the new plan according to probability P, , where T is the current temperature parameter and the initial temperature is set to 100. When the furniture layout scheme score is not improved after 10 consecutive iterations, the search is stopped and the three-dimensional coordinates and orientation parameters of the current optimal layout scheme are output and transmitted to the AR scene rendering engine and superimposed on the virtual guest room model for user preview.

[0034] A further improvement of the technical solution of the present invention is that: in the user interaction interface, the process of generating a touch and voice interaction interface for displaying VR tour results and AR customization effects includes:

[0035] Based on historical user operation data, the display priority of the interface module is dynamically adjusted. A touch and voice interaction interface is generated through a template engine, supporting gesture sliding and voice command switching of navigation and customization functions. The touch and voice interaction interface includes a navigation control area, a customization operation area, and an information display area. The navigation control area displays the VR scene switching button and path planning status. The customization operation area is embedded with a touch color wheel that displays RGB values ​​in real time, and an embedded furniture layout adjustment panel that displays translation vectors. The information display area dynamically displays basic hotel information, the RGB values ​​of matching bedding colors, and the scores of each furniture layout plan.

[0036] When a user clicks on the color wheel to select a color, the virtual hotel tour customization system captures the screen pixel coordinates, extracts the corresponding RGB values, and normalizes them to the [0, 1] range. When the user drags a furniture model, the coordinate difference between the starting and ending points is recorded and converted into an absolute displacement.

[0037] Voice signals are collected through mobile device microphones and VR headsets, and voiceprint recognition technology is used to distinguish user identities. Voice command keywords are set to trigger corresponding operations.

[0038] The optimal path node sequence output by the VR navigation module is transmitted to the interface navigation control area, and the movement path is dynamically displayed with a green line. The bedding color of the AR customization module is matched with the RGB value and furniture layout plan, and superimposed on the information display area in real time. The view rotation and zoom preview are supported. When the furniture layout score is lower than the set score threshold, a red light flashes to prompt and it is pushed to the information display area.

[0039] A further improvement of the technical solution of the present invention is that: in the data storage management module, the process of storing the hotel VR tour data and AR customization data using encryption technology includes:

[0040] The data storage management module receives VR tour and AR customization data, including the virtual hotel 3D model, path node coordinates, proxy value tables, and user movement trajectories. AR customization data includes the RGB values ​​of the user-selected bedding color, the 3D coordinates and orientation angles of the furniture layout plan, and the translation vector cluster center of historical adjustment records. It performs field-level encryption on the structured data in the VR tour and AR customization data, and uses block encryption technology for unstructured data. It combines symmetric and asymmetric encryption algorithms and generates keys based on the user's voiceprint features.

[0041] The encrypted data is stored in a database table by type. The database table includes a VR guide table and an AR customization table. The VR guide table stores the encrypted hash value of the virtual hotel 3D model, the ciphertext of the path node, and the ciphertext of the cost value. The AR customization table stores the encrypted RGB value, the ciphertext of the furniture layout plan, and the ciphertext of the user's historical operation log. The data integrity is verified by a hash function. When storing data, the ciphertext hash value is calculated and saved independently. The hash consistency is verified during access. The access rights are bound to the voiceprint identity authentication, and unauthorized users cannot decrypt the data.

[0042] After the user verifies his or her identity through voiceprint recognition, the virtual hotel tour customization system calls the pre-stored voiceprint features and regenerates the encryption key. If the key verification fails, access to the encrypted data is prohibited. When the VR tour module and AR customization module request data, the virtual hotel tour customization system decrypts the corresponding fields and data blocks. The decrypted data is temporarily stored in the memory and cleared immediately after use.

[0043] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art:

[0044] The present invention provides a virtual hotel tour customization system based on intelligent optimization and multimodal interaction. Detailed 3D spatial information of the hotel and user preferences are acquired through a multi-source data acquisition module, enabling the VR tour module to accurately construct a virtual environment and provide intelligent path planning based on collision detection, greatly enhancing the user's navigation experience and sense of immersion in the virtual hotel.

[0045] The present invention provides a virtual hotel tour customization system based on intelligent optimization and multimodal interaction. The AR customization module can intelligently adjust and optimize the room layout according to the user's customization behavior data. By calculating the scores of different layout schemes, the optimal layout scheme is automatically selected, thereby meeting the user's personalized needs and improving the customization and interactivity of the virtual tour.

[0046] The present invention provides a virtual hotel tour customization system based on intelligent optimization and multimodal interaction. The user interaction interface integrates multiple interaction methods such as touch and voice, allowing users to interact with the virtual environment more naturally. At the same time, the data storage management module adopts encryption technology to ensure the security of hotel VR tour data and AR customization data, protecting user privacy and hotel business information. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0048] Figure 1 A block diagram of the present invention. DETAILED DESCRIPTION

[0049] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0050] Examples, such as Figure 1 As shown, the present invention provides a virtual hotel tour customization system based on intelligent optimization and multimodal interaction, including a multi-source data acquisition module, a VR tour module, an AR customization module, a user interaction interface and a data storage management module;

[0051] The multi-source data acquisition module collects the hotel's 3D spatial point cloud data and user-customized behavior data, and performs pre-processing and feature extraction on them. LiDAR sensors are deployed in the center of the hotel lobby, corridors, and guest room tops, covering a 360-degree horizontal field of view at a fixed downward angle and a 60-degree vertical field of view. Depth cameras are deployed at the four corners of the guest rooms and on the columns in the public areas, 2.5 meters from the ground, with a vertical viewing angle covering the height of the room and a horizontal field of view of 90 degrees. The LiDAR emits a pulsed laser beam, measures the time difference from the emission to the reflection reception of the beam, calculates the distance to the surface of the object, and generates 3D spatial point cloud data containing three-dimensional coordinates and reflection intensity information space for each point. The depth camera uses structured light to project a pattern, captures the pattern offset after the deformation of the object surface, and combines the triangulation principle to calculate the depth value, generating dense 3D spatial point cloud data with a resolution of 0.5 cm to supplement the edge details of the furniture. The touch terminal is embedded in the hotel's official website and mobile application interface to support touch screen operation, collects the RGB values ​​of the user-selected bedding color and furniture translation instructions, and collects voice The module is integrated into the microphone of the mobile device and the VR headset, receives voice commands in real time, and distinguishes the user's identity through voiceprint recognition. When the user touches the screen to select a color block, the virtual hotel guide customization system extracts the RGB value corresponding to the screen pixel coordinates to collect the user's color selection data. When the user drags the virtual furniture model, the virtual hotel guide customization system records the coordinate difference between the starting point and the end point, obtains the layout adjustment instruction, and calculates the average distance and standard deviation of each point in the point cloud with its neighboring points based on the statistical filtering principle. If the average distance between a point and its neighboring points exceeds the average distance plus 3 times the standard deviation, it is determined to be an outlier and removed. In addition, the adjacent complete wall point cloud data is used to complete the missing wall area through plane equation fitting. In the pre-processed 3D point cloud, the path nodes of the passable area are marked and a cost value is assigned to each node. The denoised point cloud data is merged with the completed wall data to generate a 3D space model of the virtual hotel. It is input into the collision detection and path planning algorithm of the VR guide module. The cost table of the path node is transmitted to the path planning algorithm as the weight basis for calculating the moving path. The RGB value selected by the user is converted into a [0,1] interval value through normalization processing, and the starting point and the end point are combined with the room coordinate system. The coordinate differences are converted to absolute coordinates. The normalized color data and absolute coordinate layout instructions are associated with the user account and encrypted for storage. The data is then input into the AR customization module. The pre-processed 3D point cloud data is projected onto a horizontal plane and divided into grids with a side length of 0.1 meters. The number of grids covering the virtual hotel wall point cloud is counted, multiplied by the area of ​​each grid, and the total area of ​​the room is accumulated. The planar distance between the center points of each piece of furniture is calculated based on the vertex coordinates of the furniture bounding box. The length, width, and height are calculated based on the vertex coordinates of the obstacle bounding box. In the point cloud of the traversable area, the ground center point is extracted as the path node, with the spacing between adjacent nodes being 0.5 meters. Based on the area attributes of the path nodes, a cost of 1 is assigned to corridor nodes and a cost of 2 to staircase nodes. These are used as weight parameters in the path planning algorithm. The translation vectors in historical user adjustment data are counted, and the directions and magnitudes of high-frequency adjustments are calculated. The user adjustment data is classified using the K-means clustering algorithm, and the translation vectors corresponding to the cluster centers are extracted. The total room area, furniture center point spacing, and obstacle dimensions are input into the collision detection and path planning algorithms of the VR navigation module. The layout adjustment features are then fed into the AR customization module.

[0052] The VR guide module uses a collision detection algorithm based on the bounding box according to the 3D space point cloud data to calculate the position relationship between the user's virtual character position and each bounding box, and plans the optimal movement path. From the pre-processed 3D space point cloud data, the vertex coordinates of the wall and furniture are extracted, and the minimum and maximum values ​​of their three-dimensional coordinates are calculated respectively to form an axial bounding box. The user moves through the VR device. The virtual hotel guide customization system obtains the three-dimensional coordinates of the virtual character in real time based on the positioning system of the VR device, traverses the bounding boxes in the scene, and determines one by one whether the current position of the virtual character overlaps with the spatial range of the bounding box. If the x-axis, y-axis and z-axis of the virtual character's three-dimensional coordinates are all within the minimum and maximum values ​​of the three-dimensional coordinates of the bounding box, it is determined that the virtual character collides with the bounding box. If a collision is detected, the virtual hotel guide customization system executes constraints according to the bounding box type, prohibits the virtual character from entering the wall bounding box area, corrects the virtual character coordinates to the nearest edge outside the bounding box, allows the virtual character to approach the furniture bounding box area but restricts penetration, and triggers a red light flashing prompt. The bounding box types include wall and furniture The tool receives the path node coordinates, cost value table, user starting point and end point coordinate data through the VR navigation module, accumulates the cumulative cost value from the starting point to the current node according to the path node cost value, obtains the actual cost, uses the Euclidean distance to calculate the straight-line distance from the current node to the end point, obtains the estimated cost, uses the A* algorithm, defines the total cost function as the sum of the actual cost and the estimated cost, adds the starting point to the open list, records its total cost function value, and performs iterative search. The iterative search process includes selecting the node with the smallest total cost function value from the open list as the current node. If the current node is the end point, the path is backtracked and output. If the current node is not the end point, its adjacent nodes are traversed. For each adjacent node, the actual cost and the estimated cost are calculated, and the total cost function value is updated. If the adjacent node is not in the open list and there is a better path, the current node information is updated and added to the open list. The iteration is terminated until there is no feasible path, and the optimal path node sequence is output. The optimal path node sequence is transmitted to the VR scene rendering engine, displayed in real time with green lines, and the node cost value is dynamically adjusted according to the hotel traffic.

[0053] The AR customization module selects and adjusts room layout elements based on user customization behavior data, calculates the layout scores of each plan, and selects the best layout plan. It matches the normalized user-selected RGB value with the bedding color library of the virtual hotel, calculates the Euclidean distance between the user color and the colors in the bedding color library, selects the color with the smallest distance as the matching result, records its RGB value as the user preference feature, associates the RGB value of the best matching color with the user account, encrypts and stores it in the data storage management module, and overlays the matched bedding color on the virtual guest room model in the AR scene in real time. Based on the pre-processed room 3D model, furniture size data and user historical layout preference features, it generates an initial furniture layout plan, defines the furniture position using 3D coordinates and orientation angles, comprehensively analyzes the furniture spacing and distance from the wall, calculates the furniture layout plan score, applies position perturbations to random furniture in the current layout, and calculates the score of the new furniture layout plan. If the score of the new furniture layout plan is lower than the score of the current furniture layout plan , then accept the new furniture layout plan, otherwise accept the new plan according to probability P, , where T is the current temperature parameter and the initial temperature is set to 100. When the furniture layout scheme score does not improve after 10 consecutive iterations, the search is stopped and the 3D coordinates and orientation parameters of the current optimal layout scheme are output and transmitted to the AR scene rendering engine and superimposed on the virtual guest room model for user preview;

[0054] The user interaction interface generates a touch and voice interaction interface that displays VR tour results and AR customization effects. According to the user's historical operation data, the display priority of the interface module is dynamically adjusted. The touch and voice interaction interface is generated through the template engine, which supports gesture sliding and voice command switching of tour and customization functions. The touch and voice interaction interface includes a tour control area, a customization operation area and an information display area. The tour control area displays the VR scene switching button and the path planning status. The customization operation area is embedded with a touch color disk, which displays the RGB value in real time, and is embedded with a furniture layout adjustment panel to display the translation vector. The information display area dynamically displays the basic information of the hotel, the RGB value of the matching bedding color and the score of each furniture layout plan. When the user clicks the color disk to select a color, the virtual hotel tour is fixed. The system captures screen pixel coordinates, extracts corresponding RGB values, and normalizes them to values ​​in the [0,1] range. When the user drags the furniture model, the coordinate difference between the starting and ending points is recorded and converted into absolute displacement. Voice signals are collected through the mobile device microphone and VR headset, and voiceprint recognition technology is used to distinguish user identities. Voice command keywords are set to trigger corresponding operations. The optimal path node sequence output by the VR navigation module is transmitted to the interface navigation control area, and the movement path is dynamically displayed with green lines. The bedding color of the AR customization module is matched with the RGB value and furniture layout plan, and is superimposed on the information display area in real time. It supports perspective rotation and zoom preview. When the furniture layout score falls below the set score threshold, a red light flashes to indicate that the score is not higher than the set score threshold, and the information display area is pushed.

[0055] The data storage management module encrypts the hotel VR tour data and AR customization data, and uses big data storage and management technology to store the encrypted data. The VR tour data includes the virtual hotel 3D model, path node coordinates, proxy value table and user movement trajectory. The AR customization data includes the RGB value of the user's selected bedding color, the 3D coordinates and orientation angle of the furniture layout plan and the translation vector cluster center of the historical adjustment record. The structured data in the VR tour and AR customization data are encrypted at the field level. The unstructured data uses block encryption technology, combining symmetric encryption algorithm with asymmetric encryption technology, and generates a key through the user's voiceprint features. The encrypted data is stored in the database table by type. The database table includes the VR tour table and the AR customization table. The guide table stores the encrypted hash value of the virtual hotel three-dimensional model, the ciphertext of the path node and the ciphertext of the cost value. The AR customization table stores the encrypted RGB value, the ciphertext of the furniture layout plan and the ciphertext of the user's historical operation log. The data integrity is verified by the hash function. The ciphertext hash value is calculated and saved independently when storing data. The hash consistency is verified during access. The access rights are bound to the voiceprint identity authentication. Unauthorized users cannot decrypt the data. After the user verifies his identity through voiceprint recognition, the virtual hotel guide customization system calls the pre-stored voiceprint feature and regenerates the encryption key. If the key verification fails, access to the encrypted data is prohibited. When the VR guide module and the AR customization module request data, the virtual hotel guide customization system decrypts the corresponding fields and data blocks. The decrypted data is temporarily stored in the memory and cleared immediately after use.

[0056] First, the multi-source data acquisition module is responsible for collecting the hotel's 3D spatial point cloud data and the user's customized behavior data, and pre-processing and feature extraction of these data. The guest downloads and opens the software, registers and logs in to the account. Then, the user selects the VR hotel tour function on the software homepage, connects the VR device, enters the hotel's VR virtual scene, and starts a full-scale tour. The system will build a realistic virtual hotel environment based on the collected 3D data, and use a collision detection algorithm based on the bounding box to plan the optimal movement path of the user's virtual character. During the tour, the guest can pause and switch areas at any time to view the facilities and area details of interest. When the guest enters the room tour link, click the room customization button to open the mobile phone Or the camera of the tablet scans the guest room space, enters the AR scene, and switches to the AR customization mode. In the AR customization mode, the guest can select different bedding colors and furniture layouts for virtual customization by touching the screen or voice commands. By calculating the layout scores of each plan, the best layout plan is selected for the user. After the customization is completed, the guest confirms the customization plan, and the software sends the guest's customization information to the hotel backend. The user can intuitively view the VR tour results and AR customization effects through the touch and voice interaction interface, and perform interactive operations. Finally, the hotel VR tour data and AR customization data are securely stored by the data storage management module using encryption technology. After receiving the customization information, the hotel will arrange the room in advance according to the guest's requirements.

[0057] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A virtual hotel tour customization system based on intelligent optimization and multimodal interaction, characterized by: It includes multi-source data acquisition module, VR navigation module, AR customization module, user interaction interface and data storage management module; The multi-source data acquisition module collects the hotel's 3D spatial point cloud data and user customized behavior data; The VR navigation module uses a bounding box-based collision detection algorithm based on 3D spatial point cloud data to calculate the positional relationship between the user's virtual character position and each bounding box, and plans the optimal movement path; The AR customization module selects and adjusts room layout elements based on user customization behavior data, calculates the layout score of each plan, and selects the best layout plan; The user interaction interface generates a touch and voice interaction interface that displays VR tour results and AR customization effects; The data storage management module encrypts the hotel VR tour data and AR customization data, and uses big data storage and management technology to store the encrypted data; If a collision is detected, the virtual hotel tour customization system enforces constraints based on the bounding box type, prohibiting the avatar from entering the wall bounding box area, correcting the avatar's coordinates to the nearest edge outside the bounding box, allowing the avatar to approach the furniture bounding box area but restricting penetration, and triggering a red light flashing prompt. The bounding box type includes walls and furniture. In the VR navigation module, the process of planning the optimal movement path from the user's starting point to the search end point includes: Receive path node coordinates, proxy value table, user start and end point coordinate data through the VR navigation module; The actual cost is obtained by adding the cumulative cost value from the starting point to the current node according to the path node cost value. The straight-line distance from the current node to the end point is calculated using the Euclidean distance to obtain the estimated cost. The A* algorithm is used to define the total cost function as the sum of the actual cost and the estimated cost. The starting point is added to the open list, its total cost function value is recorded, and an iterative search is performed. The iterative search process includes selecting the node with the smallest total cost function value from the open list as the current node. If the current node is the end point, the path is backtracked and output. If the current node is not the end point, its adjacent nodes are traversed. For each adjacent node, the actual cost and estimated cost are calculated, and the total cost function value is updated. If the adjacent node is not in the open list and there is a better path, the current node information is updated and added to the open list. The iteration is terminated until there is no feasible path, and the optimal path node sequence is output. The optimal path node sequence is transmitted to the VR scene rendering engine, displayed in real time with green lines, and the node cost value is dynamically adjusted according to the hotel traffic.

2. The virtual hotel tour customization system based on intelligent optimization and multimodal interaction according to claim 1, characterized in that: The multi-source data acquisition module collects the hotel's 3D spatial point cloud data and user customized behavior data in the following process: LiDAR sensors are deployed in the hotel lobby, corridors, and at the center of the guest room ceilings, covering a 360-degree horizontal field of view at a fixed downward angle, with a vertical field of view of 60 degrees. Depth cameras are deployed at the four corners of the guest rooms and on pillars in public areas, 2.5 meters above the ground, with a vertical field of view covering the entire height of the room and a horizontal field of view of 90 degrees. The LiDAR emits a pulsed laser beam to generate 3D spatial point cloud data, containing three-dimensional coordinates and reflection intensity information for each point. The depth camera uses structured light projection patterns to generate dense 3D spatial point cloud data with a resolution of 0.5 cm. The touch terminal is embedded in the hotel's official website and mobile application interface to support touch-screen operation. It collects the RGB values ​​of the user's selected bedding color and furniture translation instructions. The voice collection module is integrated into the mobile device microphone and VR headset to receive voice instructions in real time and distinguish the user's identity through voiceprint recognition. When the user touches the screen to select a color block, the virtual hotel tour customization system extracts the RGB value corresponding to the screen pixel coordinates to collect the user's color selection data. When the user drags the virtual furniture model, the virtual hotel tour customization system records the coordinate difference between the starting point and the end point to obtain layout adjustment instructions.

3. The virtual hotel tour customization system based on intelligent optimization and multimodal interaction according to claim 2 is characterized by: In the multi-source data acquisition module, the process of pre-processing the collected 3D space point cloud data and user customized behavior data includes: Based on the principle of statistical filtering, the average distance and standard deviation between each point in the point cloud and its neighboring points are calculated. If the average distance between any point and its neighboring points exceeds the average distance plus three times the standard deviation, it is determined to be an outlier and removed. The adjacent complete wall point cloud data is used to complete the missing wall area through plane equation fitting. In the pre-processed 3D point cloud, the path nodes of the passable area are marked and a cost value is assigned to each node. The denoised point cloud data is merged with the completed wall data to generate a 3D spatial model of the virtual hotel. This is input into the collision detection and path planning algorithm of the VR guide module. The cost value table of the path nodes is transmitted to the path planning algorithm as the weight basis for calculating the movement path. The RGB value selected by the user is converted into a value in the range [0, 1] through normalization. The coordinate difference between the starting point and the end point is converted into absolute coordinates based on the room coordinate system. The normalized color data and absolute coordinate layout instructions are associated with the user account, encrypted and stored, and input into the AR customization module.

4. The virtual hotel tour customization system based on intelligent optimization and multimodal interaction according to claim 3 is characterized by: In the multi-source data acquisition module, the process of extracting features from the pre-processed 3D space point cloud data and user-customized behavior data includes: The preprocessed 3D point cloud data was projected onto a horizontal plane and divided into grids with a side length of 0.1 meters. The number of grids covering the virtual hotel wall point cloud was counted and multiplied by the area of ​​each grid to obtain the total area of ​​the room. The plane distance between the center points of each piece of furniture was calculated based on the vertex coordinates of the furniture bounding box. The length, width, and height were calculated based on the vertex coordinates of the obstacle bounding box. In the point cloud of the traversable area, the ground center point was extracted as the path node, with a spacing of 0.5 meters between adjacent nodes. Based on the attributes of the area where the path node is located, a cost value of 1 for corridor nodes and a cost value of 2 for stair nodes were assigned. These values ​​were used as weight parameters in the path planning algorithm. Count the translation vectors in the user's historical adjustment data, calculate the high-frequency adjustment direction and amplitude, classify the user adjustment data using the K-means clustering algorithm, and extract the translation vector corresponding to the cluster center; The total area of ​​the room, the distance between the center points of furniture, and the size of obstacles are input into the collision detection and path planning algorithm of the VR navigation module, and the layout adjustment features are input into the AR customization module.

5. The virtual hotel tour customization system based on intelligent optimization and multimodal interaction according to claim 4 is characterized by: In the VR navigation module, the process of using a bounding box-based collision detection algorithm to calculate the positional relationship between the user's virtual character position and each bounding box includes: Extract the vertex coordinates of the wall and furniture from the pre-processed 3D space point cloud data, calculate the minimum and maximum values ​​of their three-dimensional coordinates, and form an axial bounding box; The user moves through the VR device, and the virtual hotel guide customization system obtains the three-dimensional coordinates of the virtual character in real time based on the positioning system of the VR device, traverses the bounding box in the scene, and determines one by one whether the current position of the virtual character overlaps with the spatial range of the bounding box. If the x-axis, y-axis and z-axis of the virtual character's three-dimensional coordinates are all within the minimum and maximum values ​​of the bounding box's three-dimensional coordinates, it is determined that the virtual character has collided with the bounding box.

6. The virtual hotel tour customization system based on intelligent optimization and multimodal interaction according to claim 5, characterized in that: In the AR customization module, the process of matching the user-selected bedding color using a color recognition algorithm based on the RGB color space includes: The normalized RGB value of the user's selection is matched with the bedding color library of the virtual hotel. The Euclidean distance between the user's color and the colors in the bedding color library is calculated. The color with the smallest distance is selected as the matching result, and its RGB value is recorded as the user preference feature. The RGB value of the best matching color is associated with the user account, encrypted and stored in the data storage management module, and the matched bedding color is superimposed on the virtual guest room model in the AR scene in real time.

7. The virtual hotel tour customization system based on intelligent optimization and multimodal interaction according to claim 6, characterized in that: In the AR customization module, the simulated annealing algorithm is used to calculate the scores of each furniture layout plan and select the best furniture layout plan. The process includes: Generate an initial furniture layout plan based on the pre-processed 3D room model, furniture size data, and the user's historical layout preference characteristics. Define furniture positions using 3D coordinates and orientation angles, comprehensively analyze the distance between furniture and the wall, and calculate the furniture layout plan score. Apply position perturbations to random furniture in the current layout and calculate the score of the new furniture layout solution If the score of the new furniture layout plan is lower than the score of the current furniture layout plan , then accept the new furniture layout plan, otherwise accept the new plan according to probability P, , where T is the current temperature parameter and the initial temperature is set to 100. When the furniture layout scheme score is not improved after 10 consecutive iterations, the search is stopped and the three-dimensional coordinates and orientation parameters of the current optimal layout scheme are output and transmitted to the AR scene rendering engine and superimposed on the virtual guest room model for user preview.

8. The virtual hotel tour customization system based on intelligent optimization and multimodal interaction according to claim 7, characterized in that: In the user interaction interface, the process of generating a touch and voice interaction interface for displaying VR tour results and AR customization effects includes: Based on historical user operation data, the display priority of the interface module is dynamically adjusted. A touch and voice interaction interface is generated through a template engine, supporting gesture sliding and voice command switching of navigation and customization functions. The touch and voice interaction interface includes a navigation control area, a customization operation area, and an information display area. The navigation control area displays the VR scene switching button and path planning status. The customization operation area is embedded with a touch color wheel that displays RGB values ​​in real time, and an embedded furniture layout adjustment panel that displays translation vectors. The information display area dynamically displays basic hotel information, the RGB values ​​of matching bedding colors, and the scores of each furniture layout plan. When a user clicks on the color wheel to select a color, the virtual hotel tour customization system captures the screen pixel coordinates, extracts the corresponding RGB values, and normalizes them to the [0, 1] range. When the user drags a furniture model, the coordinate difference between the starting and ending points is recorded and converted into an absolute displacement. Voice signals are collected through mobile device microphones and VR headsets, and voiceprint recognition technology is used to distinguish user identities. Voice command keywords are set to trigger corresponding operations. The optimal path node sequence output by the VR navigation module is transmitted to the interface navigation control area, and the movement path is dynamically displayed with a green line. The bedding color of the AR customization module is matched with the RGB value and furniture layout plan, and superimposed on the information display area in real time. The view rotation and zoom preview are supported. When the furniture layout score is lower than the set score threshold, a red light flashes to prompt and it is pushed to the information display area.

9. The virtual hotel tour customization system based on intelligent optimization and multimodal interaction according to claim 8, characterized in that: In the data storage management module, the process of using encryption technology to store hotel VR tour data and AR customization data includes: The data storage management module receives VR tour and AR customization data, including the virtual hotel 3D model, path node coordinates, proxy value tables, and user movement trajectories. AR customization data includes the RGB values ​​of the user-selected bedding color, the 3D coordinates and orientation angles of the furniture layout plan, and the translation vector cluster center of historical adjustment records. It performs field-level encryption on the structured data in the VR tour and AR customization data, and uses block encryption technology for unstructured data. It combines symmetric and asymmetric encryption algorithms and generates keys based on the user's voiceprint features. The encrypted data is stored in a database table by type. The database table includes a VR guide table and an AR customization table. The VR guide table stores the encrypted hash value of the virtual hotel 3D model, the ciphertext of the path node, and the ciphertext of the cost value. The AR customization table stores the encrypted RGB value, the ciphertext of the furniture layout plan, and the ciphertext of the user's historical operation log. The data integrity is verified by a hash function. When storing data, the ciphertext hash value is calculated and saved independently. The hash consistency is verified during access. The access rights are bound to the voiceprint identity authentication, and unauthorized users cannot decrypt the data. After the user verifies his or her identity through voiceprint recognition, the virtual hotel tour customization system calls the pre-stored voiceprint features and regenerates the encryption key. If the key verification fails, access to the encrypted data is prohibited. When the VR tour module and AR customization module request data, the virtual hotel tour customization system decrypts the corresponding fields and data blocks. The decrypted data is temporarily stored in the memory and cleared immediately after use.

Images